Machine Learning Applications in Movement Biomechanics

Dr. Marion Mundt

This tutorial will provide an overview of different applications of machine learning to support movement biomechanics analysis. You will learn to use an off-the-shelf pose estimation model to determine anatomically-related landmarks in 2D videos and to code a simple artificial neural network to classify normal and abnormal gait from inertial sensor data. The pros and cons of machine learning in biomechanics will be discussed and the opportunities and challenges explained using the two examples. All analyses will be undertaken using samples Python files.

Biography

Marion is a Research Fellow in the UWA Tech & Policy Lab at The University of Western Australia, working with the Australian Institute of Sport to use machine learning techniques to estimate motion parameters from standard two-dimensional video. She received her PhD in Sports Science from the German Sport University Cologne for the application of artificial intelligence to motion analysis. Using inertial sensors, health-related information like joint loads can be collected for different movements during daily life situations, enabling the collection of big data in biomechanics on-the-fly.